Congestive heart failure (CHF) poses an enormous medical, societal and financial burden in the U.S, and hospitalization accounts for the majority of costs. Over 40% of hospitalized CHF patients have diabetes, and treatment of hyperglycemia during acute illness has been found to improve morbidity and mortality in other disease states. However, the relationship between congestive heart failure exacerbation and improved glycemic control is not known. Studies report that acute ("stress") hyperglycemia, but not chronic hyperglycemia predicts CHF mortality, suggesting that acute glycemic instability may be important. The investigators therefore wish to determine prospectively whether glycemic variability is related to outcomes during CHF exacerbation. Glycemic variability is associated with more profound endothelial toxicity than tonic glucose elevations in vitro, and in patients with diabetes, it is associated with oxidative stress and ischemic EKG changes, independent of sustained hyperglycemia. Glycemic variability has also been independently associated with ICU mortality, but otherwise, it has not been studied prospectively in any hospitalized patient population. The primary aims of this study are to determine whether CHF patients undergoing continuous intravenous insulin therapy have improved glycemic variability relative to a physiologic subcutaneous insulin regimen, and to investigate whether glycemic variability leads to improvement in established prognostic variables in patients with CHF. Specifically, the study will examine heart rate variability, markers of inflammation, endothelial function, BNP, oxidative stress, disease severity, and quality of life scores in patients with CHF exacerbation. This protocol will be conducted among 80 patients admitted to the Ohio State University Ross Heart Hospital with hyperglycemia and decompensated heart failure due to systolic dysfunction (ejection fraction <35%). Patients will be randomly assigned to one of 2 groups: (1) intravenous insulin therapy targeting a blood glucose of 150 mg/dL, (2) subcutaneous insulin with basal, prandial, and supplemental components. Glucose will be checked hourly in all patients. Intravenous insulin will be continued for 72 hours. The results will be applied to larger patient populations in order to examine hard clinical endpoints. PUBLIC HEALTH RELEVANCE: Despite advances in medical care for heart failure, hospitalizations and deaths are rising, and patients with diabetes are at increased risk. The current study investigates whether fluctuations in blood sugar (glucose) during deterioration of heart failure play a role.